Calculate the price of Items in stored in list form

Question

I have a Dataframe which looks like this:Item-table

Date.    Item.     
10-sep.  X,Y,Z
11-sep.  Y,Z
12-sep.  Z
13-sep.  Z,X

And another Table where price of each item is stored date wise. Price-table

Item.   10sep.  11sep.   12sep.  13sep
X.       10.     5.        10.      15
Y.        7.     15.       13.       10
Z.        5.      10.       10.      10

I want my output to look like this:

Date.   Item.    Total Price
10 sep.  X,Y,Z.   22
11 sep.  Y,Z.     25
12 sep.  Z.       10
13 sep.  Z,X.     25

In first row total ptice is 22 because Price of X,Y and Z on 10 sep is 10,7 and 5 respectively. May i know how i can get this output column.

Answer 1

I am going to use this dataframes to solve your problem

print(df1)
     Date          Item      
0  10-sep         X,Y,Z 
1  11-sep           Y,Z 
2  12-sep             Z 
3  13-sep           Z,X 

print(df2)
  Item     10sep    11sep     12sep    13sep
0    X        10        5        10       15
1    Y         7       15        13       10
2    Z         5       10        10       10

We can use DataFrame.lookup to select the values of the data frame 2, but first we must prepare the values to do the search:

df3=df1.copy()
df3['Item']=df3['Item'].str.split(',')
df3=df3.explode('Item')
df3['Date']=df3['Date'].str.replace('-','')
print(df3)

    Date Item
0  10sep    X
0  10sep    Y
0  10sep    Z
1  11sep    Y
1  11sep    Z
2  12sep    Z
3  13sep    Z
3  13sep    X

---

mapper=df2.set_index('Item')


print(mapper)
      10sep  11sep  12sep  13sep
Item                            
X        10      5     10     15
Y         7     15     13     10
Z         5     10     10     10

---

df3['value']=mapper.lookup(df3['Item'],df3['Date'])
df1['Total Price']=df3.groupby(level=0).value.sum()
print(df1)
     Date          Item  Total Price
0  10-sep         X,Y,Z           22
1  11-sep           Y,Z           25
2  12-sep             Z           10
3  13-sep           Z,X           25

---

Time comparison for this dataframes:

method of Valdi_Bo:

%%timeit
ItemPrice = Prices.set_index('Item').stack().swaplevel().rename('Price')
def totalPrice(row):
    dat = row.Date
    items = row.Item.split(',')
    ind = pd.MultiIndex.from_arrays([[dat] * len(items), items])
    return ItemPrice.reindex(ind).sum()
Items['Total Price'] = Items.apply(totalPrice, axis=1)
13.5 ms ± 699 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

my method

%%timeit
df3=Items.copy()
df3['Item']=df3['Item'].str.split(',')
df3=df3.explode('Item')
mapper=Prices.set_index('Item')
df3['value']=mapper.lookup(df3['Item'],df3['Date'])
Items['Total Price']=df3.groupby(level=0).value.sum()
7.68 ms ± 178 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

---

@anky_91 method

%%timeit
m=df2.set_index('Item').T
n=df1[['Date']].assign(**df1['Item'].str.get_dummies(',')).set_index('Date')
final=df1.set_index('Date').assign(Total_Price=m.mul(n).sum(1)).reset_index()
8.7 ms ± 199 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

Answer 2

I assumed some minimal order and coordination between your both DataFrames, ie:

• There are no trailing dots is column names.
• Date format in column names in Prices is just like in Date column in Items (they can be of string type, but both of them have a minus after the day number.

So Items and Prices Dataframes are actually as follows:

     Date   Item
0  10-sep  X,Y,Z
1  11-sep    Y,Z
2  12-sep      Z
3  13-sep    Z,X

  Item  10-sep  11-sep  12-sep  13-sep
0    X      10       5      10      15
1    Y       7      15      13      10
2    Z       5      10      10      10 

The first step is to convert Prices into a Series :

ItemPrice = Prices.set_index('Item').stack().swaplevel().rename('Price')

so that it contains:

        Item
10-sep  X       10
11-sep  X        5
12-sep  X       10
13-sep  X       15
10-sep  Y        7
11-sep  Y       15
12-sep  Y       13
13-sep  Y       10
10-sep  Z        5
11-sep  Z       10
12-sep  Z       10
13-sep  Z       10
Name: Price, dtype: int64

Then define a function to compute a total price:

def totalPrice(row):
    dat = row.Date
    items = row.Item.split(',')
    ind = pd.MultiIndex.from_arrays([[dat] * len(items), items])
    return ItemPrice.reindex(ind).sum()

And the last step is to apply this function to each row and save the result as a new column:

Items['Total Price'] = Items.apply(totalPrice, axis=1)

The result is:

     Date   Item  Total Price
0  10-sep  X,Y,Z           22
1  11-sep    Y,Z           25
2  12-sep      Z           10
3  13-sep    Z,X           25

Answer 3

Taking the cleaned data courtesy @Valdi_Bo, you can also try get dummies and multiply with the transposed dataframe and sum on axis=1 to get your desired output:

m=df2.set_index('Item').T
n=df1[['Date']].assign(**df1['Item'].str.get_dummies(',')).set_index('Date')
final=df1.set_index('Date').assign(Total_Price=m.mul(n).sum(1))

---

print(final)

         Item  Total_Price
Date                      
10-sep  X,Y,Z           22
11-sep    Y,Z           25
12-sep      Z           10
13-sep    Z,X           25